Chimpanzee_vis
Scott Kum
November 16, 2020
Rationale
Cancer immunotherapy is a breakthrough strategy for cancer treatment. However, its utility has been limited to few highly immunogenic cancer types. Therefore, many recent research efforts have been focused on improving immune targeting of the so-called “cold” tumors. One such effort involve vaccinating patients with neoantigens expressed by cancer cells, but so far, only the exome, which maximally cover 2% of the human genome, has been explored for these vaccines. I aim to harness transposons (te) which cover up to 50% of the human genome, to develop personalized cancer vaccine to improve cancer immunotherapy efficiency and tumor specificity (Waldman, Fritz & Lenardo, 2020).
Load Libraries
tidyverse, viridis, and plotly are loaded.
Import Data
Three tsv files are imported:
ht29_gene <- read.table("HT29_HERVK_GENE.tsv", sep = "\t", header = TRUE)
ht29_te <- read.table("HT29_HERVK_TE.tsv", sep = "\t", header = TRUE)
gene_conv <- read.table("gene_conv.tsv", sep = "\t", header = TRUE)
HT29_HERVK_GENE.tsv
This is a list of read-gene pairs from a Nanopore cDNA sequencing data of HT29 cell line mapped to a repeat-masked human genome.
head(ht29_gene)## read refseq
## 1 00074061-c794-4ec8-97e2-3ee0f3886e64 NM_015659.3
## 2 000957f3-d210-4abf-9c91-7fe58162ddb6 NM_001256856.1
## 3 000d6a96-8500-4755-8f73-ddad65b607ea NM_001044387.2
## 4 00216fd0-1527-469e-bb0c-4669c275866e NM_001144072.2
## 5 002bbb2c-b6c9-47e0-9d0a-4ba35f361dd0 NM_058182.4
## 6 003b6f07-e752-4372-b409-5674e11e0863 NM_014691.3
HT29_HERVK_TE.tsv
This is a list of read-te pairs from a Nanopore cDNA sequencing data of HT29 cell line mapped to a human te consensus sequences.
head(ht29_te)## read te
## 1 00074061-c794-4ec8-97e2-3ee0f3886e64 ALU
## 2 000957f3-d210-4abf-9c91-7fe58162ddb6 ALU
## 3 000d6a96-8500-4755-8f73-ddad65b607ea ALU
## 4 00216fd0-1527-469e-bb0c-4669c275866e ALU
## 5 002bbb2c-b6c9-47e0-9d0a-4ba35f361dd0 ALU
## 6 003b6f07-e752-4372-b409-5674e11e0863 ALU
gene_conv.tsv
This is a list of RefSeq accesion-gene symbol pairs.
head(gene_conv)## refseq gene
## 1 NM_000014.5 A2M
## 2 NM_000015.3 NAT2
## 3 NM_000016.5 ACADM
## 4 NM_000017.4 ACADS
## 5 NM_000018.4 ACADVL
## 6 NM_000019.4 ACAT1Convert Gene Name
RefSeq accessions found in ht29_gene is not quite human-readable. The code below will replace these accessions with gene symbols based on gene_conv.
ht29_gene <- ht29_gene %>%
inner_join(gene_conv, by = "refseq") %>%
select(read, gene)
head(ht29_gene)## read gene
## 1 00074061-c794-4ec8-97e2-3ee0f3886e64 RSL1D1
## 2 000957f3-d210-4abf-9c91-7fe58162ddb6 BTRC
## 3 000d6a96-8500-4755-8f73-ddad65b607ea ZNF557
## 4 00216fd0-1527-469e-bb0c-4669c275866e UBAC2
## 5 002bbb2c-b6c9-47e0-9d0a-4ba35f361dd0 SMIM11A
## 6 003b6f07-e752-4372-b409-5674e11e0863 AQRGet Chimeric Gene-TE Pairs
inner_join is again used to join the common reads between ht29_gene and ht29_te, and the columns “gene” and “te” are selected.
chimera_pair <- inner_join(ht29_gene, ht29_te, by = "read") %>%
select(gene, te)
head(chimera_pair)## gene te
## 1 RSL1D1 ALU
## 2 BTRC ALU
## 3 ZNF557 ALU
## 4 UBAC2 ALU
## 5 SMIM11A ALU
## 6 AQR ALUCount Chimeric Gene-TE Pairs
To count the occurence of each gene-te pair, chimera_pair is grouped and counted, and then filtered to exclude outliers. complete is used to fill in missing combinations of gene and te.
chimera_count <- chimera_pair %>%
group_by_all() %>%
count() %>%
ungroup() %>%
filter(n > 6) %>%
filter(n < 50) %>%
complete(gene, te)
head(chimera_count)## # A tibble: 6 x 3
## gene te n
## <chr> <chr> <int>
## 1 BRI3BP ALU 10
## 2 BRI3BP MER34B_I NA
## 3 BRI3BP UHG NA
## 4 CD24 ALU 30
## 5 CD24 MER34B_I NA
## 6 CD24 UHG NAVisualize Data as Heatmap
A heatmap was chosen to visualize this data because it allows discrete values in both x- and y-axis, and further allows a third dimension of quantity (column “n” in chimera_count) to be represented with color. ggplot2 is used to generate a heatmap with genes along the x-axis and te along the y-axis. Variety of theme options are used to make the plot look better.
ggplot(data = chimera_count, mapping = aes(x = gene, y = te, fill = n)) +
geom_tile(colour = "white", size = 1.5, stat = "identity") +
coord_equal() +
xlab("") +
ylab("") +
ggtitle(label = "HT29-HERVK Chimeric Pairs \n") +
theme_light() +
theme(axis.text.x = element_text(size = 30, angle = 90, vjust = 0.5, hjust=1),
axis.text.y = element_text(size = 30),
legend.key.height = unit(0.5, "inch"),
legend.key.width = unit(1, "inch"),
legend.text = element_text(size = 30),
legend.position = "right",
legend.title=element_blank(),
plot.title = element_text(size = 40)) +
scale_fill_viridis(option = "A", na.value = "gray")
ggplotly is used in conjunction with ggplot2 for an interactive heatmap experience. Few theme options used above are modified or omitted here due to incompatibility with the ggplotly function.
Interpretation
From the above heatmap, ALU element seems to form a chimera with a wide variety of genes. Since ALU element is present in the 5’ and 3’ UTR of many genes, some of those chimeric reads may not be a unique feature of cancer. However, ALU element is still capable of mobilization and so a thorough look at the position of these ALU elements within genes is warranted.
Although it may look as if the diversity of gene-te chimera is somewhat limited, the starting data is a 50-fold random subset of the dataset from a single Nanopore cDNA sequencing run, and therefore application of this pipeline to the full dataset will likely yield more pairs than seen here. This will also likely lead to a clearer picture of which chimera pair is truly enriched in the transcriptome of HT29 cells, an important factor in selecting a candidate for cancer vaccine.
Future Direction
One major drawback of the analysis pipeline described above is that it assumes a 1-to-1 relationship between gene and te within a chimeric transcript. However, chimera involving more than two of these elements have been observed, albeit rarely. It is difficult to visually represent such chimera though, and ggplot may not be suitable for these edge cases.
One important aspect of this research is to identify unique chimera signatures between cancer cell lines and ultimately between patients to allow personalized medicine. Therefore, I will additionally sequence the transcriptome of other human cancer cell lines and further develop this pipeline to compare data between samples.